Abstract Fault geometries dictate the spatiotemporal distribution of seismicity and govern earthquake propagation, serving as the foundation for seismic hazard assessment. Recent advances in seismological monitoring have significantly lowered the detection threshold of local-scale earthquakes, providing an opportunity to delineate complex fault geometries from a seismological aspect. However, the current extraction and interpretation of such lineaments mostly rely on human examination, which might be subjective and time-consuming. This study proposes an objective method to automatically detect and parameterize complex linear structures (i.e., faults and fractures) from extensive earthquake catalogs. Taking source parameters into consideration, Earthquake to Fault (E2F) contains three significant steps: (1) using the 3D Hough transform to identify the linear features, (2) clustering the fault candidates with a fault width-based clustering radius, and (3) extracting the final fault parameters with three evaluation parameters (i.e., utilization rate of seismic event, 3D aspect ratio, and ratio of seismic moment [RSM]). We successfully decoupled and extracted >61 linear features from the 21,617 events (−2.19<Mw<3.21) recorded at a hydraulic fracturing site (i.e., Tony Creek Dual Microseismic Experiment), in which both complex fracture networks and kilometer-long faults with variable orientations are present. We also extended E2F to various complex settings (ranging from local-scale induced to large-scale tectonic events) and successfully mapped the conjugate faults and intersecting subfaults. We believe that E2F could serve as a handy tool for interpreting and comparing earthquake catalogs and could benefit further applications that require detailed fault geometries, such as geomechanical modeling and physics-based hazard assessments.